Fault diagnosis of non-linear dynamic processes using identified hybrid models

This work addresses an approach for fault diagnosis of industrial processes using identified hybrid models. This paper concerns the identification of the hybrid model parameters through the input-output data acquired from the non-linear process. In order to show the effectiveness of the developed technique, the results obtained in the fault diagnosis of an industrial plant are finally reported.